Small ubiquitin-like modifier (SUMO1-3) is a group of proteins that conjugate to lysine residues of target proteins, and thereby modulate various processes that play important roles in key cellular functions under normal and pathological conditions. These functions include gene expression and genome integrity, proteasomal degradation of proteins, protein quality control, and DNA-damage repair (Hay 2005; Heun 2007; Prudden et al. 2007; Bergink and Jentsch 2009; Geoffroy and Hay 2009; Tatham et al. 2011). In the intact brain, levels of SUMO2/3-conjugated proteins are very low, but are massively increased after transient cerebral ischemia and during deep hypothermia (Lee et al. 2007; Cimarosti and Henley 2008; Cimarosti et al. 2008; Yang et al. 2008b, c). The post-ischemic activation of SUMO2/3 conjugation is believed to be an endogenous neuroprotective stress response, as neurons are extremely sensitive to even a very short period of ischemia-like conditions when SUMO2/3 expression is silenced (Datwyler et al. 2011).
Brain damage caused by transient cerebral ischemia is of significant clinical relevance, as it is a complication of major cardiovascular surgeries in pediatric and adult patients. To protect organs from ischemic damage, major cardiovascular operations requiring a period of circulatory arrest are usually performed during deep hypothermia. Depending on the clinical center where these procedures are performed, patients are exposed to various degrees of deep or moderate hypothermia, ranging from 16°C to 30°C (Jacobs et al. 2001; Arnaoutakis et al. 2007; Kamiya et al. 2007; Camboni et al. 2008; Apaydin et al. 2009; Khaladj et al. 2009; Numata et al. 2009; Takami et al. 2009; Elefteriades 2010).
Using a rat model of deep hypothermic cardiopulmonary bypass (CPB), we have demonstrated that deep hypothermia at 18°C massively activates the SUMO conjugation pathway (Yang et al. 2009). In this study, we investigated whether more moderate hypothermia is sufficient to activate SUMO conjugation. Because the potential of manipulating the SUMO conjugation pathway for preventive and therapeutic purposes is of tremendous clinical interest, we evaluated the effects of hypothermia on SUMO conjugation using an animal model of CPB and exposed animals to normothermic (37°C) or hypothermic (30°C, 24°C, or 18°C) CPB.
Here, we demonstrated that deep hypothermia triggered a shift in immunoreactivity of the SUMO-conjugating enzyme, Ubc9, from the cytoplasm to the nucleus of neurons. Furthermore, moderate hypothermia at 30°C was sufficient to induce a marked rise in levels and nuclear accumulation of SUMO2/3-conjugated proteins in neurons. Because various pathways modulated by SUMO conjugation are nuclear processes pivotal for cells to recover from stress, as discussed above, and as mounting evidence suggests that activation of SUMO conjugation increases the tolerance of cells to stress conditions (Lee et al. 2007, 2009, 2011; Yang et al. 2009; Datwyler et al. 2011), it is conceivable to conclude that SUMO conjugation plays a role in the organ protection that hypothermia provides.
- Top of page
- Materials and methods
- Supporting Information
Recently, using the same animal model as in this study, we demonstrated that exposing rats to 1 h of 18°C deep hypothermic CPB markedly activates SUMO2/3 conjugation, and to a lesser extent, SUMO1 conjugation (Yang et al. 2009). This effect was observed in the hypothermic CPB animals, but not in the normothermic CPB group, indicating that the surgical procedure per se does not cause the activation (Yang et al. 2009). Here, we exposed animals to CPB at varying temperatures ranging from 18°C to 30°C, i.e, the full range of deep to moderate hypothermia used clinically for major cardiovascular surgeries, as discussed above. Specifically, rats were subjected to 37°C, 18°C, 24°C, or 30°C CPB for 1 h, and changes in levels of SUMO-conjugated proteins were analyzed by Western blot. A moderate increase in SUMO1 conjugation and a marked increase in SUMO2/3 conjugation occurred in the hypothermic CPB animals (Figure S1; Fig. 1a and b). Levels of SUMO1-conjugated proteins rose 2.7 ± 0.6-fold, 3.8 ± 1.0-fold, and 4.1 ± 0.7-fold in the cortex, 3.5 ± 0.7-fold, 4.2 ± 0.8-fold, and 3.8 ± 0.7-fold in the hippocampus, and 2.8 ± 0.8-fold, 3.3 ± 0.4-fold, and 2.7 ± 0.4-fold in the striatum of animals exposed to 30°C, 24°C, or 18°C CPB, respectively, compared to normothermic 37°C CPB (Figure S1). Changes in levels of SUMO2/3-conjugated proteins were more pronounced, increasing 11.1 ± 3.1-fold, 17.8 ± 5.5-fold, and 17.1 ± 6.3-fold in the cortex, 5.7 ± 3.2-fold, 8.0± 1.5-fold, and 5.6 ± 2.8-fold in the hippocampus, and 10.8 ± 3.0-fold, 15.4 ± 3.3-fold, and 13.0 ± 4.5-fold in the striatum of animals exposed to 30°C, 24°C, or 18°C, respectively, compared to normothermic 37°C CPB (Fig. 1). These data indicate that SUMO2/3 conjugation was activated even after moderate hypothermia of 30°C.
Figure 1. Marked activation of small ubiquitin-like modifier (SUMO)2/3 conjugation during moderate to deep hypothermia. Western blot analysis depicts the pattern and summary of SUMO2/3 conjugation in the brains of animals subjected to normothermic (37°C) or 1 h of moderate to deep hypothermic (30°C, 24°C, or 18°C) cardiopulmonary bypass (CPB; a, b), or 0 min 18°C deep hypothermic CPB (c). Monoclonal antibody against β-actin was used as loading control. The high-molecular-weight area in each lane, as indicated in (a), was cropped and analyzed. Quantification data are presented as means ± SD (n = 5/group). **p < 0.01; ***p < 0.001 (anova, followed by Fisher's protected least-significant difference (PLSD) test).
Download figure to PowerPoint
In the first set of experiments, CPB animals were cooled down to the target temperature over a period of 30 min, and brains were processed 1 h after animals had reached the target temperature. To determine whether SUMO2/3 conjugation is activated after a very short exposure to deep hypothermia, CPB animals were cooled down to 18°C, and brains were sampled immediately when animals reached the target temperature, i.e, 30 min after starting the cooling process. Levels of SUMO2/3-conjugated proteins rose 9.5 ± 1.8-fold, 5.7 ± 1.2-fold, and 7.0 ± 1.0-fold in the cortex, hippocampus and striatum, respectively (Fig. 1c).
A large portion of SUMO-conjugation targets are nuclear proteins involved in various pathways that play key roles in the recovery of stressed cells. Thus, for the activation of SUMO2/3 conjugation to be of key functional significance, we would expect it to occur preferentially in nuclei of neurons. To investigate the relevance of hypothermia in this process, CPB animals were cooled down to 18°C, 24°C, or 30°C, brains were sampled when animals had reached the target temperature, and changes in SUMO2/3 conjugation were evaluated by immunohistochemistry. As activation of SUMO2/3 conjugation in the brain is predominantly a neuronal stress response (Yang et al. 2008b), sections were costained for the neuron-specific cytoplasmic marker MAP2. In normothermic brains, SUMO2/3 immunoreactivity was mainly confined to the cytoplasm of neurons, both in the cortex and the hippocampal CA1 subfield (Fig. 2a and b). However, in hypothermic animals, strong SUMO2/3 immunoreactivity was present in nuclei and markedly less in the cytoplasm of neurons.
Figure 2. Massive nuclear accumulation of small ubiquitin-like modifier (SUMO)2/3-conjugated proteins in cortical and hippocampal neurons exposed to even moderate hypothermia. Representative immunofluorescence staining shows the pattern of SUMO2/3 conjugation in the cortex (a) and hippocampus (b) of animals subjected to normothermic (37°C) or moderate to deep hypothermic (30°C, 24°C, or 18°C) cardiopulmonary bypass (CPB). Microtubule-associated protein (MAP)2 staining (red) was used to identify neurons. Scale bar: 10 μm.
Download figure to PowerPoint
To determine whether the nuclear SUMO2/3 immunoreactivity found in hypothermic animals represented free SUMO2/3, conjugated SUMO2/3, or both, we separated brain extracts into cytosolic and nuclear fractions. In nuclear fractions isolated from both 37°C normothermic and 18°C hypothermic animals, we did not find free SUMO2/3 (Figure S2). This suggests that nuclear SUMO2/3 immunoreactivity found in neurons of hypothermic animals (Fig. 2a and b) represented SUMO2/3-conjugated proteins and not free SUMO2/3.
To evaluate the mechanisms underlying the observed increase in nuclear SUMO2/3 immunoreactivity in hypothermic animals, we stained brain sections of 37°C normothermic and 18°C hypothermic animals for Ubc9, the only SUMO-conjugating enzyme identified so far, and MAP2 as neuronal cytoplasm marker. In normothermic brains, Ubc9 immunoreactivity was mainly confined to the cytoplasm of neurons, both in the cortex and hippocampal CA1 subfield (Fig. 3a and b), similar to the pattern described by Lee et al. (2011). However, in 18°C hypothermic animals, Ubc9 immunoreactivity was also present in the nuclei and, clearly less, in the cytoplasm of neurons. This suggests that Ubc9 nuclear translocation contributes to the marked increase in levels of SUMO2/3-conjugated proteins found in the nuclei of neurons of hypothermic animals.
Figure 3. Nuclear translocation of Ubc9 immunoreactivity in cortical and hippocampal neurons of animals exposed to 18°C hypothermia. Representative immunofluorescence staining shows the pattern of Ubc9 in the cortex (a) and hippocampus (b) of animals subjected to normothermic (37°C) or hypothermic (18°C) cardiopulmonary bypass. Microtubule-associated protein (MAP)2 staining (red) was used as a cytoplasmic marker of neurons. Scale bar: 10 μm.
Download figure to PowerPoint
In search of a cell culture model for investigating the mechanisms underlying activation of SUMO2/3 conjugation under deep hypothermic conditions, we used primary neuronal cultures. Human neuroblastoma SHSY5Y cells and primary cortical neurons pre-exposed to 4°C deep hypothermia are more tolerant to transient OGD (Lee et al. 2007; Loftus et al. 2009). We exposed primary neuronal cell cultures, prepared from cortices of embryonic rat brains, to deep hypothermia. As activation of SUMO2/3 conjugation is a response of cells to various kinds of stresses (Yang et al. 2008a), we used MAP2 immunostaining to identify hypothermia-induced morphological changes as an indicator of stressed cells following hypothermia exposure (Fig. 4). Cells exposed to 4°C, but not 16°C, showed marked dendrite degeneration including beading formation and fragmentation (Fig. 4a). We therefore focused on 4°C deep hypothermia for further analyses. Next, cells were exposed to 4°C for 5, 15, or 30 min (Fig. 4b). We observed degenerative changes in cells exposed to prolonged 4°C hypothermia (Fig. 4b, 30 min).
Figure 4. Morphological damage, indicative of cellular stress, illustrated by dendritic degeneration following prolonged exposure of primary neuronal cell cultures to 4°C deep hypothermia. Primary neuronal cultures were prepared from cortices of embryonic rat brains. Cultures were exposed to 16°C or 4°C hypothermia for 30 min or 2 h (a), or to 5, 15, or 30 min 4°C hypothermia (b). Mean arterial blood pressure (MAP)2 (red), and DAPI (blue) staining were used to identify neurons and nuclei, respectively.
Download figure to PowerPoint
To elucidate the effects of deep hypothermia on SUMO2/3 conjugation, neuronal cultures were exposed to 4°C deep hypothermia with or without normothermic recovery (Fig. 5). A moderate rise in levels of SUMO2/3-conjugated proteins occurred during a short exposure to 4°C deep hypothermia (Fig. 5a). However, the change in the pattern of SUMO2/3 conjugation was only transient, as it was almost normalized after only 5 min of exposure. Activation of SUMOylation was rapidly reversed when cells were transferred back to 37°C normothermic conditions (Fig. 5b).
Figure 5. Small transient rise in levels of small ubiquitin-like modifier (SUMO)2/3-conjugated proteins in primary neurons exposed to 4°C hypothermia. Western blot analysis depicts the pattern (a, b) and summary (c, d) of SUMO2/3 conjugation in primary neuronal cultures exposed to 4°C hypothermia without (a, c) or with (b, d) normothermic recovery. Monoclonal antibody against β-actin was used as loading control. The high-molecular-weight area in each lane, as indicated in (a,b), was cropped and analyzed. Quantification data are presented as means ± SD (n = 3/group). ***p < 0.001; a, cp < 0.05, p < 0.001 versus 0 min (0 min, c) or 0-min recovery (0 min rec., d), respectively (anova, followed by Fisher's protected least-significant difference (PLSD) test).
Download figure to PowerPoint
In the last set of experiments, cells were subjected to 30-min OGD and 24 h of recovery with or without prior deep hypothermia exposure. Hoechst/PI staining indicated that transient OGD-induced severe cell death that was partially suppressed in cultures exposed to 4°C deep hypothermia for 5 min, but not for 15 min, 2 h prior to normothermic OGD (Fig. 6a). When the interval between hypothermia and OGD was extended to 24 h, this protective effect vanished (Fig. 6b). We quantified the extent of cell death following OGD by counting Hoechst- and PI-positive cells. Cell death was suppressed by about 25% after a short 5-min period of deep hypothermia 2 h prior to OGD (Fig. 6c).
Figure 6. Acute but not delayed tolerance to oxygen/glucose deprivation (OGD) following a short period of 4°C hypothermia. Primary neurons were exposed to 5- or 15-min 4°C hypothermia (37°C, control) for 2 h (a) or 24 h (b) prior to 30-min normothermic transient OGD and 24-h recovery. Representative immunofluorescence staining depicts the pattern of nuclei (Hoechst staining) and damaged cells (PI staining), respectively. Quantification data (c) are presented as means ± SD (n = 3/group). **p < 0.01 hypothermia versus normothermia (anova, followed by Fisher's protected least-significant difference (PLSD) test).
Download figure to PowerPoint
- Top of page
- Materials and methods
- Supporting Information
The potential of deep hypothermia to protect organs from damage induced by transient ischemia is well established. However, the mechanisms underlying organ protection by deep hypothermia and strategies to maximize its efficacy still need to be established. Understanding these mechanisms would be a pivotal step toward designing therapeutic strategies to activate these processes and thus induce a state of tolerance to transient ischemia without risking the adverse effects associated with deep hypothermia. The most promising strategy would be to activate endogenous neuroprotective pathways before performing surgical procedures that require a period of circulatory arrest, and thus increase the resistance of neurons to a transient interruption of blood supply. The SUMO conjugation pathway could be such a pathway, which, when activated, protects neurons from ischemic insults.
A massive activation of SUMO conjugation induced by deep hypothermia was first reported by John Hallenbeck and his colleagues (Lee et al. 2007). Using hibernating squirrels as an experimental model to identify endogenous neuroprotective stress-response pathways, they found massive activation of SUMO conjugation during the state of hibernation torpor (Lee et al. 2007). During torpor, when the body temperature of animals is sharply reduced to about 5°C, blood flow, energy consumption, and protein synthesis are lowered to otherwise lethal levels (Frerichs et al. 1994, 1998; Frerichs and Hallenbeck 1998; Carey et al. 2003). The investigators therefore hypothesized that SUMO conjugation plays a role in tolerance to the severe ischemia associated with hibernation torpor. This hypothesis was substantiated by results of experiments on cell cultures exposed to transient oxygen/glucose deprivation (OGD, ischemia-like conditions) (Lee et al. 2007, 2009; Datwyler et al. 2011). Furthermore, using transgenic mice over-expressing exogenous Ubc9, the only SUMO-conjugation enzyme identified to date, a recent study showed that increased SUMO conjugation was associated with an increased tolerance to transient focal cerebral ischemia (Lee et al. 2011).
The most prominent finding of this study is that exposure to moderate 30°C hypothermia in rats was sufficient to markedly modify the pattern and the subcellular localization of SUMO2/3 in neurons, and that hypothermia triggered nuclear translocation of the SUMO-conjugating enzyme Ubc9 (Figs 1-3). In brains of animals exposed to normothermic CPB, most of the SUMO2/3 immunoreactivity was detectable as free SUMO (Fig. 1, band at about 18 kDa), and strong SUMO2/3 immunoreactivity was present in the cytoplasm but not in the nuclei of neurons (Fig. 2; 37°C). In animals exposed to hypothermic CPB, however, levels of SUMO2/3-conjugated proteins increased markedly (Fig. 1), and strong SUMO2/3 immunoreactivity was present in nuclei but not in the cytoplasm of neurons (Fig. 2). This suggests a massive nuclear accumulation of SUMO2/3-conjugated proteins in neurons of hypothermic animals.
It is well established that the speed of chemical and biochemical reactions is temperature-dependent, a decrease in temperature retarding the reactive process. One plausible explanation for the neuroprotective effects of hypothermia, therefore, is that it protects cells from ischemic damage by retarding the rate of energy depletion during ischemia. Hypothermia does indeed depress the tricarboxylic acid flux (Kaibara et al. 1999) and thus preserve cerebral energy metabolism during ischemia (Laptook et al. 1995; Yager and Asselin 1996; Williams et al. 1997), resulting in delayed anoxic depolarization (Kaminogo et al. 1999). These observations imply that hypothermia-induced neuroprotection is a passive process whereby the rate of glucose metabolism is lowered, consequently delaying the time to terminal depolarization. This would shorten the period of terminal depolarization during transient ischemia and thus mitigate all pathological processes triggered during the state of energy depletion and manifested after recovery from ischemia.
On the other hand, transient hypothermia induces rapid and delayed forms of tolerance to ischemic injury, suggesting that it can trigger an active process that protects cells from damage caused by ischemia (Nishio et al. 2000; Yunoki et al. 2002, 2003). Given that SUMO2/3 conjugation protects neurons from ischemia-like conditions (Datwyler et al. 2011), that hypothermia provides neuroprotection, and that hypothermia triggers nuclear translocation of Ubc9 and activation of nuclear SUMO2/3 conjugation (Figs 2 and 3), we conclude that SUMO contributes to the neuroprotective process induced by hypothermia.
In search of a simplified experimental model to investigate, in future studies, the mechanisms underlying the deep hypothermia-induced rise in levels of SUMO2/3-conjugated proteins and to screen for drugs that activate this process in the absence of hypothermia, we cultured primary neurons from embryonic rat brains. Exposure to deep hypothermia induced only a short-lasting very moderate increase in levels of SUMO2/3-conjugated proteins (Fig. 6). This suggests that neurons in culture have a much lower capacity for activating SUMO2/3 conjugation than neurons in the intact brain. This notion is supported by our recent observation that transient OGD induced only a moderate rise in levels of SUMO2/3-conjugated proteins in primary neuronal cultures (Datwyler et al. 2011). However, SUMO2/3 conjugation was dramatically increased in post-OGD B35 neuroblastoma cells (Yang et al. 2012) and also in neurons in vivo subjected to a transient interruption in blood supply (Yang et al. 2008a, b). This indicates that the capacity of the SUMO conjugation machinery to respond to stressful conditions is much lower in primary neurons than in established cell lines in culture, and much less in cultured neurons than in neurons in the intact brain.
Various factors may contribute to this obvious discrepancy, including the developmental state of neurons and the microenvironment. The expression levels of SUMO2/3-conjugated proteins and the SUMOylation machinery are temporally regulated in the developing brain (Loriol et al. 2012), which may explain why primary neurons in culture isolated from embryonic brains react to hypothermic or metabolic stress differently than neurons in the adult brain. Furthermore, neurons are post-mitotic non-dividing cells that survive in vivo for the entire lifespan of the animal, but only for a limited time when they are isolated from brains and kept in culture. It would be interesting to determine whether the viability of neurons in culture is limited by inactivity of the SUMOylation machinery.
The microenvironment may also contribute to the different response of in vivo versus in vitro neurons to both deep hypothermia and transient ischemia/OGD. In the intact brain, neurons are physically and functionally tightly associated with glial cells, and this mutual interaction is important for physiologic function. In vitro, culture conditions are usually chosen that suppress growth of glial cells, resulting in very pure neuronal cultures. Furthermore, neurons are surrounded by a very small extracellular space in vivo, whereas neurons in culture are bathed in an ocean of medium. Elucidating the mechanisms underlying the limited ability of cultured neurons to respond to stress conditions by massively activating SUMO2/3 conjugation may help to better understand the role of SUMO conjugation in protecting cells from damage.
Many of the SUMO2/3 conjugation target proteins identified so far are transcription factors and other nuclear proteins that modify gene expression and play key roles in DNA-damage repair. We therefore expect the massive nuclear accumulation of SUMO2/3-conjugated proteins in neurons of hypothermic animals to contribute to the neuroprotective effects of hypothermia. Deep hypothermia does indeed modify gene expression and down-regulates expression of several genes that are associated with the pathological process triggered by transient ischemia (Yang et al. 2009). These include chemokines, intracellular adhesion molecule-1, interferon regulatory factor-1, and interleukin-1β and interleukin-6 (Yang et al. 2009).
Recently, an increasing number of SUMOylation target proteins have been identified in neurons that are not nuclear but rather cytosolic or cell membrane proteins (Martin et al. 2007b; Wilkinson et al. 2010). These include the mRNA-binding protein La (van Niekerk et al. 2007), the kainate receptor subunit GluR6 (Martin et al. 2007a), and the potassium channel Kv1.5 (Benson et al. 2007). This suggests a role for SUMO conjugation in modulating axonal mRNA trafficking, excitability of neurons, and ion fluxes (Wilkinson et al. 2010). Whether SUMOylation of cytosolic or cell membrane proteins is activated during hypothermia and contributes to the protective effect must be determined in future studies. We observed that during hypothermia, most of the SUMO2/3 immunoreactivity is present in the nuclei of neurons (Fig. 2). This supports the notion that nuclear processes mediated by SUMO2/3 conjugation are activated during hypothermia.
The rise in levels of SUMO2/3-conjugated proteins in neurons in the intact brain exposed to deep hypothermia or transient ischemia is much more dramatic than in cultured neurons exposed to deep hypothermia or transient ischemia-like conditions, as discussed above and demonstrated recently (Datwyler et al. 2011). In vivo studies are therefore needed to verify whether the activation of SUMO conjugation during hypothermia or after ischemia is indeed a protective stress response, and to establish strategies to manipulate the SUMO conjugation pathway for preventive and therapeutic purposes. Transient hypothermia/ischemia experiments on SUMO-knockout animals will provide important information on the role of individual SUMO paralogues in protecting neurons from ischemic damage. SUMO-transgenic animals may help us to identify proteins that are SUMO2/3-conjugated in neurons of hypothermic brains. Such investigation will be integral to unraveling the mechanisms by which SUMO conjugation imparts neuroprotection during hypothermia. This will be pivotal to identify drugs that activate these processes in normothermic animals, thus avoiding the potentially adverse effects associated with deep hypothermia.